The changing world of tape - Storage As I See It - Industry Overview - Column

Computer Technology Review, August, 2002 by Mark Ferelli

This month's issue of Computer Technology Review takes an especially close look at small form factor tape, now and in the not to distant future. It is always chancy looking at this area, because products and vendors are constantly in motion. The players change: Exabyte and Ecrix have merged, MaxOptix has gone into the autoloader business under the name of Peak, and AIWA has exited the autoloader market outright.

One of the other fairly constant changes is in the expectations of end users. There is an evident shift to higher capacities, performance characteristics and pricing that will likely continue well into the decade. Yes, higher prices for tape products. But that doesn't mean that end users are going to be less price-sensitive. Quite the contrary. But the product mix is changing, and there is a price impact.

The very-lowest-end products are seeing an end of life. They just can't keep up with disk drive capacities and throughput speeds. Even more significant, low cost tape drives are being replaced by alternative technologies, most significantly writable compact disks. Additionally, although DAS is still the dominant peripheral architecture in the marketplace, there is a clear migration to network-centric storage, recentralized IT management, and a new recognition of the independent value of stored data to the enterprise.

Bottom line, there is more data to be stored at faster speeds than ever before. Disk rotational speeds are hitting 15,000rpm. Transfer rates are in the megabytes/second and more. The result: The enterprise will require higher capacities, higher performance characteristics, and consequently more expensive tape drives. Curiously, tape drive manufacturers will continue to be under the gun to drive down prices for individual drive models while the average drive price is rising.

Part of the changing world of tape is reflected in the methods used by tape vendors, for product improvement. If you examine many of the roadmaps for tape migration for some of the tape technologies we cover at CTR, you'll see that the cost-effective areas for improvement are centered in heads and media. There are some excellent formulations in media right now...the sophistication of materials chemistry at all nine media vendors is exceptional. New formulations allow more data per square inch of tape. But improvements are not limited to chemistry. The concept of mid-point load that first appeared in the IBM Magstar made a significant cut at tape latencies. The memory-in-cassette (MIC) chip that is found in Sony's AIT cartridges has eased the problems of finding files...another chop at latency times.

Lots of motion in the tape area from the smallest cartridges to the largest silos. Significant R&D investment exists at the drive, media and automation levels. Increasingly sophisticated software for managing backups, recovery, archiving and hierarchical storage management. As network storage grows in disk capacities and transfer rates, tape will necessarily follow. Now, who was trying to tell me about tape's upcoming and inevitable death? What's the punch line?